This month Twitter rolled out a polling functionality to all users. With just two options and running for only 24 hours, the polls are very simplistic and are very unlikely to set the world of market research alight.
But as proponents of more agile and nimble ways of conducting research we were intrigued to find out how and indeed whether this new tool worked. So we put it to the test.
We set up a poll from the FreshMinds Twitter account, asking our followers to vote on whether they preferred Halloween or Bonfire Night- a tough choice for many, myself included! And we decided to ask the same question on Google Consumer Surveys to see how the two tools compared.
We’re testing how Twitter polls compare to Google Surveys. Vote & we'll share our findings. Which do you prefer?
— Decidedly (@DecidedlyHQ) November 3, 2015
Interestingly both surveys returned broadly similar results. Bonfire Night came out on top, gaining 59% of the vote on Twitter, 63.6% on Google Surveys. But in the course of the experiment we learnt much more than that consumers prefer Bonfire Night , we discovered some of the positives and drawbacks of both tools.
The good: Twitter polls provide a means of engaging with and gaining feedback from two extremely valuable audiences. Firstly, consumers who are interested in and passionate about your brand i.e. your followers. Secondly, when when combined with popular hashtags, it enables you to reach and poll users who are interested in particular topics. A great example of this came from Marketing Magazine who just last week tapped into the hugely popular #onthemoon hashtag to understand reactions to the new John Lewis Christmas advert:
Do you think this year's John Lewis Christmas advert #OnTheMoon was better than last year's Monty the penguin?
— Campaign: Brands (@CampaignBrands) November 6, 2015
The bad: Twitter, by its very nature, encourages short-form communication. But this can be problematic when it comes to using the social networking platform to poll your audience. When making use of the polling feature, Twitter’s already small character limit is reduced even further to a mere 114 characters, meaning that questions have to be short, and as a result, simplistic. Likewise, with the ability to add only two possible answers, the polling feature has a tendency to be somewhat reductive.
The ugly: Did you know the life of a tweet is just 18 minutes? I didn’t but it’s a stumbling block I came up against as part of this experiment. Ensuring you can collect enough responses to your poll before it disappears, buried under the mass of other tweets, is really difficult! It requires a lot of time and effort, retweeting and linking to your original tweet, to secure participation.
Because of this, obtaining a large enough sample size can be difficult. Despite having a following of almost 6, 000, we were only able to secure 20 responses to our poll. Marketing Magazine also came up against similar obstacles with their John Lewis poll: only 206 of their 322, 000 followers took part. In contrast, Google Consumer Surveys, with its ability to recruit a far greater number of participants (reaching into the thousands) is far more robust.
Google Consumer Surveys
The good: When it comes to targeting, Google Consumer Surveys is streets ahead of Twitter. Whilst on Twitter you are at the whim of whoever chooses to answer your poll, Google’s tool allow you to target respondents by age, gender and location. What’s more, you are then able to analyse your results by these variables, which can prove useful. For instance, on our poll we were able uncover that a greater percentage of men prefer Bonfire Night to Halloween than women.
The bad: Whilst Google Consumer Surveys enables targeting, there are limits to this. The age and gender of respondents is not known, and instead it is inferred, based on the individual’s browsing behaviour and the location of their IP address, which is then combined with income and urban density data to make an educated guess. For me, and many other insight professionals, this means that the quality of the data Google provides is just not robust enough.
The ugly: Google Consumer Surveys are served up to two different types of respondents: the first are users of the Google Opinion Rewards App who receive Google Play credits in return for completing surveys: in this way working in a similar way to many research panels. The second group of respondents are web users who have to complete the surveys in order to access premium online content such as articles. And it’s this latter group of respondents that really worries me. I know that if I was in their position I’d try to click through the survey as quickly as possible, giving no thought to the questions or indeed my responses, which could seriously compromise the insight uncovered. Given our commitment to ‘keeping it real’ and trying to get to the heart of consumers’ actual behaviour, it’s this drawback to Google Consumer Surveys that really vexes us!
Due to some of the complications we’ve discussed, I am yet to be convinced of the true value of either Google Consumer Surveys or Twitter Polls in their current format.
With an ability to reach a large, broad-based sample at minimal cost, Google Consumer Surveys does show at least some potential. Perhaps, going forwards it could have a role to play in helping organisations quickly poll broad audiences to ascertain attitudes towards particular topics? But Twitter Polls – at least for the time being – are limited to the novelty value of being able to survey a small proportion of your followers.
Whilst neither Google Surveys or Twitter Polls are yet able to provide big brands with insights that are robust, tailored and rich enough to form the basis of their most significant business decisions, we do need to avidly follow further developments in this space. Not only from Google, and Twitter, but from other new players looking to break into the market research industry.